Add edges vector to CodeInstance/CodeInfo to keep backedges as edges
#54894
+705
−769
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
vtjnash
force-pushed
the
jn/codeinfo-edges
branch
from
af946a0
to
0277c82
Compare
vtjnash
force-pushed
the
jn/codeinfo-edges
branch
2 times, most recently
from
af03fd5
to
91be20b
Compare
vtjnash
force-pushed
the
jn/codeinfo-edges
branch
from
91be20b
to
ce5726f
Compare
|
@nanosoldier |
|
Your benchmark job has completed - possible performance regressions were detected. A full report can be found here. |
vtjnash
force-pushed
the
jn/codeinfo-edges
branch
from
ce5726f
to
d2c2e51
Compare
|
Given that this is now a pure regression size-wise, can the description be updated to more clearly explain why this PR is a good thing? |
vtjnash
force-pushed
the
jn/codeinfo-edges
branch
3 times, most recently
from
2f824bf
to
33701f1
Compare
|
I'm working on rebasing this. |
vtjnash
force-pushed
the
jn/codeinfo-edges
branch
from
33701f1
to
4525784
Compare
aviatesk
force-pushed
the
jn/codeinfo-edges
branch
3 times, most recently
from
a7aae3c
to
c83f06d
Compare
|
@nanosoldier |
|
Your benchmark job has completed - possible performance regressions were detected. A full report can be found here. |
aviatesk
force-pushed
the
jn/codeinfo-edges
branch
from
c83f06d
to
bfed3fe
Compare
|
@nanosoldier |
|
Your benchmark job has completed - possible performance regressions were detected. A full report can be found here. |
aviatesk
force-pushed
the
jn/codeinfo-edges
branch
3 times, most recently
from
a9c4eeb
to
fe8d042
Compare
This records all invoke targets as edges as a functionality test, before finishing the implementation of recording the edges accurately during inference (via backedges + inference).
Start computing edges from stmt_info later (after CodeInstance is able to have been allocated for recursion) instead of immediately.
To avoid calling `add_edges!` directly. In fact, it might be better to define something like `VirtualizedCallInfo` rather than using `MethodResultPure`.
aviatesk
force-pushed
the
jn/codeinfo-edges
branch
from
fe8d042
to
a5a2db9
Compare
aviatesk
force-pushed
the
jn/codeinfo-edges
branch
from
a5a2db9
to
6f64fd6
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Appears to add about
1MB3MB (117MB to 120MB) to the system image, and to decrease the stdlib size by2MB54 MB (305MB to 251MB), so seems overall favorable right now. The edges arenotbeing computedyetfollowing the encoding https://hackmd.io/sjPig55kS4a5XNWC6HmKSg?both#Edges-Encoding to correctly reflect the backedgesbut they currently are just extracting the.:invokefieldsTODO: we currently store 2 copies of the edges during serialization (in both the svec and the old format) which should help recover a portion of the overhead for the stdlibs.TODO: this records edges only to the MI, not to the recursive edges (as required for correctness)